Abstract
Background and goals While Zinfandel berries often ripen asynchronously, creating heterogeneity in berries from the same cluster, the effect on grape and wine attributes is not fully known.
Methods and key findings Zinfandel berries from two vineyard sources were segregated into 10-, 12-, and 14-mm berries; raisins; and unsorted berries. These were assessed for physical and chemical parameters and made into wine over three vintages. The 14-mm berries and raisins accounted for 41% to 48% and 7% to 10% of the fresh weight (FW) distribution, respectively. The largest variation across berry sizes was 41%, in 2019. Total soluble solids and ethanol levels were lowest in 14-mm berries, and the highest ethanol levels in wines were produced from 10-mm berries. Wine color, total phenolics, tannins, non-tannin phenolics, and polymeric pigments were inversely related to berry size. Wines from unsorted, 12-, and 14-mm berries had higher acidity and red fruit aromas, and lower aromatic intensity. Wines from 10-mm berries had more spicy aromas, whereas raisins resulted in aromas and flavors such as cooked fruit, raisins, cough syrup, black fruit, and enhanced sweetness.
Conclusions and significance Berries from the smallest size categories were the main drivers of elevated ethanol levels in Zinfandel wines. Most color and phenolic parameters in wines were linked with the skin surface-to-berry FW ratio in the berries, which was highest in 10-mm berries and lowest in unsorted and 14-mm berries. Smaller berries were responsible for the trademark spicy and dried-fruit, raisin-like character of Zinfandel wines.
Introduction
The Vitis vinifera L. grape berry is composed of the pericarp (skin), the mesocarp (pulp), and seeds (0 to 4 per berry). Berry size is genetically controlled (Gray and Coombe 2009), although environmental factors such as natural or imposed water deficit, and practices such as the choice of rootstock (Ziegler et al. 2020) and irrigation (Ojeda et al. 2001), also have profound effects on berry volume and weight. Berry size varies greatly within different V. vinifera varieties, but differences are also often found within the same variety, vine, and cluster (Pagay and Cheng 2010), including variations in the proportion of seeds, skins, and pulp. This proportion is considered a key factor affecting the final chemical and sensory composition of the wines. For example, on a fresh weight (FW) basis, Cabernet Sauvignon berries are on average 5% seed, 15% skin, and 80% pulp (Roby and Matthews 2004); meanwhile, most solutes of sensory relevance are located within the solids of the berry.
Although final berry size is the result of the cell number multiplied by the mean cell size for each tissue, differences in berry size within different V. vinifera varieties are mostly a function of cell enlargement, rather than of the rate of cell division (Fernandez et al. 2006). The implication is that smaller berries within a given variety are smaller because of comparatively smaller (yet not less abundant) cells. This implication is also expected to hold true for intrinsic variations in berry size within the berries of a single cluster or variety, which can be as high as 60% among berries from a single plot (Singleton 1972).
Berry size has significant implications for winemaking. While the domestication of most crops targeted increased fruit sizes (Cong et al. 2008), this does not apply to winegrapes. Indeed, from an agronomical point of view, the selection for smaller berry size in V. vinifera winegrapes, which are generally used for winemaking purposes, relative to, for example, table grapes, is linked to the notion that smaller berries have less water and thus higher concentration of solutes. This claim has been proven unfounded (Matthews and Nuzzo 2007). However, winemakers report that smaller berries seem to have higher concentrations of solutes of sensory relevance, producing higher quality wines (Jackson 2014). These solutes include phenolic compounds (such as anthocyanins, tannins, and other phenolic classes), their reaction products formed during winemaking, and free and bound volatile compounds that are eventually released and/or transformed into wine aromas. The skin tissue contains anthocyanins, which are pigments primarily responsible for wine color, and tannins, which are phenolic polymers of variable molecular weight that are responsible for textural sensations such as astringency (Casassa and Harbertson 2014). Seeds also contain tannins (albeit of lower molecular weight and at higher concentration than skin-derived tannins), which contribute to wine astringency, and monomeric flavan-3-ols, which are bitterants (Monagas et al. 2003). Free and bound volatile compounds, while also present in the pulp, are mainly located in the skins and include methoxypyrazines (vegetal aromas), terpenes (floral aromas), nor-isoprenoids, and thiols (González-Barreiro et al. 2015), among others. Because these solutes are primarily confined to the solid components of the berry (i.e., skins and seeds), it is assumed that smaller berries should have a higher solid-to-liquid ratio, and thus are deemed more suitable for winemaking.
The effect of berry size on berry physical and chemical composition has been widely investigated, but only a few studies have extended this to the resulting wines. For example, within six berry size categories in Cabernet Sauvignon, skin-derived anthocyanins, tannins, and total soluble solids (TSS) decreased with increasing berry size (Roby et al. 2004). Different pruning methods resulted in variations in size of Cabernet Sauvignon grapes, with smaller berries having higher concentrations of anthocyanins and total phenolics (Holt et al. 2010). Another study of Cabernet Sauvignon segregated berries into small (≤0.75 g), medium (0.76 to 1.25 g), and large (>1.25 g) size categories. In the berries, total skin mass, TSS, anthocyanins, and total phenolics decreased with increasing berry size, while larger berries had higher malic acid and pH. In the wines, medium-size berries had the highest anthocyanins and total phenolics, but the wines made from smaller berries had the highest color saturation (Chen et al. 2018). Similar results were also reported in Cabernet Sauvignon from Spain: out of three berry size categories, smaller berries had higher proportion of skins, lower proportion of seeds, higher acidity, lower pH, and higher TSS, also producing wines with the highest color saturation (Gil et al. 2015).
Differences in berry size resulting from the irrigation practice known as regulated deficit irrigation were separated as those arising from concentration of solutes (due to reduction of berry size) and those ascribed to biosynthesis (enhanced amount of a given solute on a per berry basis). Enhanced biosynthesis of skin tannins was observed in the larger berries that received the most irrigation, whereas the smaller berries, which received much less irrigation, had higher concentrations of anthocyanins and seed tannins, ultimately resulting in wines with the highest color saturation and polymeric pigment content (Casassa et al. 2016). Finally, another study selected two separate vineyards of Carménère known to produce different berry sizes (85.7 and 133 mL, volume of 100 berries). In the resulting wines, the final extraction of phenolic compounds and volatiles such as esters and terpenes were enhanced in the berries of larger volume, which was linked to these berries having a higher skin mass proportion (skin thickness) than the smaller berries (Gil Cortiella et al. 2020). Thus, the latter study implies that skins are in fact a tissue with volume and content rather than a surface. Overall, an informal meta-analysis of the previous studies emphasizes that within a single variety, the smaller berries within the berry size distribution have distinct and more desirable characteristics for winemaking. A special challenge of past research on berry size lies in the fact that most studies produced limited quantities of wine and thus sensory analysis was not possible.
The present study was undertaken over three consecutive vintages covering two old-vine (80- to 101-yr-old) Zinfandel vineyards located in the Paso Robles American Viticultural Area (AVA) of the Central Coast of California. Zinfandel is known for its asynchronous ripening, creating heterogeneity not only within vines of the same plot, but especially within berries of the same cluster, as previously shown in other varieties (Mucalo et al. 2015). The presence of raisined berries is also common. Despite this, there is no work to date that documents this putative heterogeneity within Zinfandel berries, nor is it clear if different berry sizes within Zinfandel clusters, including raisins, have specific and unique physical, chemical, and sensory features. The aim of the present work was to document the physical and chemical composition of grapes, as well as the sensory consequences on finished wines produced from selected berry size categories, in Zinfandel from the Central Coast of California. A sorting method for the grape berry sizes was designed to sort enough berries for wine production to accommodate sensory analysis.
Materials and Methods
Fruit sources, harvest determination, and vineyard details
This 3-yr study was conducted in the Central Coast of California, San Luis Obispo County. V. vinifera L cv. Zinfandel grafted on Vitis rupestris Scheele cv. St. George rootstock from two different vineyards located in the Paso Robles AVA was used for the experiment. The vineyard source used in 2017 (35°34′N; 120°42′W), hereby referred to as “Dusi vineyard”, was planted in 1945 in the Templeton Gap District AVA of Paso Robles. The soil is composed of gravelly alluvial material, classified as sandy loam texture, with a pH of 7.3 and a calcium content of 73.8%, relative to the base saturation measurements. The vineyard plot has a minor slope (<5%; estimated using the Clinometer application for IOS [Plaincode]) and faces southeast. Row orientation is north-south, with other vineyard details previously reported (Riffle et al. 2022). The vineyard utilized in 2018 (35°33′N; 120°45′W), hereby referred to as “Pesenti vineyard”, was planted in 1924. The soil is of sedimentary origin, classified as sandy loam texture, with a pH of 7.2 and a high calcium content (83.1%). The vineyard plot is in a steep slope (~35%; estimated using the Clinometer application) and faces north. Row orientation is north-south. Vines from both vineyards were head-trained and spur-pruned, with two buds per spur. Canopies were suckered in April each year, with lateral shoot thinning and leaf thinning applied to the fruiting zone each year in July. Both sites were dry-farmed, with the vines spaced 2.4 m × 2.4 m, and farmed strictly according to the same viticultural practices. Approximately 650 kg of fruit was hand-harvested during each harvest, from Dusi vineyard in 2017 and from Pesenti vineyard in 2018 and 2019. Harvest dates were scheduled when the average TSS concentration of juice samples that were obtained from a composite vineyard sample reached between 23 and 24 Brix (±0.5; measured using a digital density meter [Anton Paar]), and when pH was above 3.2 (±0.1; determined using a benchtop pH meter [Thermo Fisher Scientific]). Immediately after harvest, the fruit was transported to the research winery at California Polytechnic State University San Luis Obispo and refrigerated at 5.5 ± 0.1°C for 12 hr prior to processing.
Seasonal growing data
All annual and seasonal weather reports, including annual and seasonal precipitation (mm) and daily high, low, and average temperatures (°C), were obtained from the California Irrigation Management Information System (CIMIS), station #163, located in Atascadero, CA (35°28′N; 120°38′W). The weather station is located 11.86 and 13.16 km away from Dusi and Pesenti vineyards, respectively. Growing degree days were calculated based on daily average temperatures from 1 April to 31 Oct of each year (base 10°C).
Berry sorting methodology, experimental design, and winemaking
In 2017, berries were destemmed by hand. In 2018 and 2019, berries were gently destemmed using a Bucher crusher-destemmer (Bucher Vaslin North America); to avoid crushing and juice release, the crushing rollers were fully disengaged. Visual inspection of the berries upon machine destemming confirmed that their integrity was directly comparable to that of hand-destemmed berries. Only intact and not crushed berries were amenable to sorting. These berries were sorted by diameter using an in-house designed device consisting of a set of aluminum trays containing 14-, 12-, and 10-mm diam holes, with the trays arranged and vertically set in place 10 cm apart from each other on an aluminum-framed structure (Supplemental Figure 1). Berries larger than 14 mm were collected in the 14-mm tray, whereas berries smaller than 10 mm were collected in a bottom tray with 8-mm holes and subsequently identified and classified as raisins, based on size and visual aspect. Sorting occurred as follows: berries were placed (~1 kg at a time) on the top tray (16 mm), and with a manual horizontal constant movement, the berries fell through each sieve, subsequently trapped according to berry diameter (mm) in the corresponding tray. After shaking each individual tray, berries were emptied into separate plastic bins for each size category and kept in a refrigerated room, with the sorting process taking place within 1 day.
From each of the four size categories obtained after sorting (raisins, 10-, 12-, and 14-mm), as well as the control (i.e., unsorted berries), 9 ± 0.1 kg of fruit was added to separate 12-L plastic fermentors (Spiedel), with triplicate fermentations (n = 3) established for each berry size category. Each fermentor received a 3-min mechanical punch-down with a plastic plunger, followed by an addition of 50 mg/L SO2. Five hours after SO2 addition, each fermentor was inoculated with commercial dry yeast (EC-1118, Saccharomyces cerevisiae, Lallemand) at a rate of 35 g/hL, following manufacturer instructions for yeast rehydration. Approximately 48 hr after crush, each fermentor was inoculated with commercial malolactic bacteria at a rate of 1 g/hL (VP-41, Oenococcus oeni, Scott Laboratories). All fermentors received two 30-sec manual punch-downs per day during alcoholic fermentation, at 0800 and 1700 hr. Temperature and TSS were monitored daily during alcoholic fermentation after each punch-down, using a basic portable density meter (Anton Paar), and fermentation temperatures were held between 25 and 27°C by placing the fermentors in a temperature-controlled room. In all cases, the wines were pressed 14 days postinoculation using a 20-L stainless steel bladder press (Spiedel) set to a pressure of 4 bars. Each wine was pressed into 3.785-L glass carboys and left to settle in a room with controlled temperature (between 18 and 20°C) until the completion of malolactic fermentation. Malic and lactic acids were monitored enzymatically using a Y15 analyzer (Biosystems). Not all treatments completed malolactic fermentation at the same time and some wines did not begin malolactic fermentation, so the wines were allowed to complete malolactic fermentation for ~10 wk postpressing. After this period, all wines were racked from their lees and adjusted to 25 mg/L free SO2. The wines were topped to avoid ullage, placed into a walk-in cooler set at 5°C, and allowed to cold-settle and mature for 5 wk. The wines were then re-racked, adjusted to a molecular SO2 level of 0.3 mg/L, and bottled into 375-mL dark green glass bottles (Encore) with DIAM 5 micro-agglomerated cork closure (G3 Enterprises). The bottled wines were kept at a constant cellar temperature of 14°C until analysis.
Berry sampling and analysis
Three replicates of 30 berries (n = 3) from each size category were randomly selected for berry physical analysis, which included berry FW; seeds per berry; FW of seeds, skins, and pulp (g); berry skin surface (cm2); and dry weight (DW) of skins and seeds (g). After weighing, individual berries were peeled using a metal spatula. The skin from each berry was blotted using a paper towel, removing any remaining pulp, extended flat on a dry piece of paper towel to remove excess moisture, and weighed. The seeds from each berry were removed from the pulp and cleansed using a paper towel, then counted and weighed. Skin surface was obtained using an optical area meter (LI-COR 3100). Both skins and seeds from each sample were placed on aluminum dishes and dried using a convection oven (Thermo Fisher Scientific) at 60°C until constant weight was obtained (3.5 hr for seeds and 5 hr for skins). Because of the difficulty of effectively removing the skins from the pulp and measuring their surface, the raisin-size category was discarded for the physical analysis.
Additional samples were taken from each size category, including the control (i.e., unsorted berries), for berry chemical analyses. Approximately 120 berries (n = 3) from each size category were randomly selected and placed into plastic sealable bags. The samples were manually crushed to release the must and allowed to macerate for ~6 hr in a walk-in cooler set at 5.5 ± 0.1°C. Analysis of selected chemical parameters of the resulting juices was conducted using a Y15 analyzer (Biosystems). Commercial enzymatic kits (Biosystems) were used for analysis of malic acid, lactic acid, tartaric acid, glucose + fructose, ammonia, alpha-amino compounds as N, yeast assimilable nitrogen, and potassium. Titratable acidity (TA) was measured by titrating a sample of juice (5 mL) in a water solution against a 0.067 N NaOH solution, following an established protocol (Iland et al. 2012). TSS for each berry size category were measured as previously described.
A total of 10 kg of destemmed, unsorted berries from each vineyard was used to determine the distribution of FW among the different size categories. The berries were sorted by size following the sorting procedure previously detailed, and each of the four obtained size categories were weighed. The percentage of FW relative to the 10 kg was calculated for each size category.
Wine basic chemical composition
Wine TA and pH were measured following the same method detailed above for the determination of juice TA and pH. Ethanol (% v/v) was measured by near-infrared spectroscopy using an Alcolyzer Wine M/ME analysis system (Anton Paar). Acetic acid, glucose, fructose, malic acid, and lactic acid were determined enzymatically using commercial enzymatic analysis kits (Admeo, Biosystems Group).
Color and phenolic analyses
The spectrophotometric measurements included analysis of phenolic compounds and chromatic parameters (wine color and hue). For all three vintages, the wines were analyzed at pressing (day 14) and at 260 days postcrush. The wines of the 2017 vintage were reassessed after 6 yr of bottle aging. Prior to analysis, wine samples were centrifuged at 15,000 × g in a microfuge (model 5415D; Eppendorf), and the supernatant was transferred into clean 1-mL Eppendorf tubes prior to analysis. Anthocyanins and total polymeric pigments (herein defined as the sum of small polymeric pigments [SPP] and large polymeric pigments [LPP]) were measured as previously reported (Harbertson et al. 2003). Tannins were analyzed by protein precipitation (Harbertson et al. 2002). Total phenolics, also referred to as “iron-reactive phenolics”, were determined by reaction with ferric chloride (Harbertson et al. 2003). Wine color was determined by placing an aliquot of undiluted wine samples in 1-mm path-length quartz cuvettes, and calculated as the sum of absorbances at 420, 520, and 620 nm. Wine hue was calculated as the ratio of absorbances at 420 and 520 nm (Glories 1984) and reported as values obtained under these conditions (1-mm path-length quartz cuvettes). All spectrophotometric measurements were made with a Cary 60 UV–Vis spectrophotometer equipped with an 18-sample cell auto-sampler (Agilent Technologies).
Sensory analysis
The wines of all three vintages were analyzed using generic descriptive analysis following standard procedures previously described (Lawless and Heymann 2010). The wines of the 2018 and 2019 vintages were analyzed after 3 mo of bottle aging, whereas the wines of the 2017 vintage were analyzed after 1 yr of bottle aging. Four skilled tasters with extensive wine tasting experience performed a pre-evaluation of all the wines before sensory analysis to confirm the absence of off-odors or reduction aromas. For the wines of the 2017 and 2018 vintages, the trained panel was composed of nine individuals (n = 9; two females and seven males) between 23- and 40-yr-old. This panel evaluated the 2017 and 2018 wines in two separate panels. Panelists were previously screened for visual disorders using Ishihara maps, and bitterness sensitivity to 6-n-propylthiouracil (PROP) (Fluka Chemical Company) was screened using a test that allows the identification of a condition known as PROP status (Tepper et al. 2001). The results obtained indicated that none of the panelists had preexisting color deficiencies, and the panel was composed of 33% supertasters, 45% medium tasters, and 22% non-tasters.
For each of the two panels, panelists were trained during a total of six 75-min sessions. Terminology was developed and determined through panel consensus, defining a total of seven aroma attributes (earthy, dark fruit, red fruit, dried fruit, cough medicine, floral, and spice), six color components (low saturation, medium saturation, high saturation, brown, red, and purple hue), and three mouthfeel attributes (acidity, sweetness, and astringency). A 10-cm unstructured scale was used for all descriptors during the training evaluations. The line scales did not have wording; indents placed at 1 cm and 9 cm from the origin represented low and high intensity, respectively. All aroma attributes and brown, red, and purple reference standards were prepared at high levels only, representing the 10 cm or highest point of the line scale (Supplemental Table 1). Per panel request, mouthfeel and saturation standards were prepared at the low, medium, and high levels, representing the 0-, 5-, and 10-cm marks on the line scale, respectively. All initial standards were presented to the panelists during the first three training sessions after terminology development; final standards were modified according to panelist feedback and consensus and presented during the final three training sessions.
For both vintages (2017 and 2018), the berry size categories representing wines made from raisins; 10-, 12-, and 14-mm berries; and unsorted berries were analyzed by the panel during four formal evaluation sessions held after the training sessions were finalized. The sessions took place in the sensory facility of the Food Science Department in the San Luis Obispo campus of the California Polytechnic State University. Each wine, including all three replicates, was presented to the panelists monadically. Measured 30-mL aliquots of wine were served at room temperature (18°C) in ISO wine tasting glasses covered with a convex glass to trap volatiles. A total of 15 wines (5 treatments × 3 replicates) were assessed per session, yielding a total of 540 observations for each vintage (15 wines × 9 panelists × 4 sessions). The wine glasses were randomly assigned three-digit codes and randomly presented to each panelist. The randomized blocks used for the evaluations were designated by a sensory analysis software, which also recorded the ratings for each attribute (Red Jade Sensory Solutions). Panelists were provided with individual iPad devices (Apple) for each session, and scores were registered digitally through the software and transferred into an Excel spreadsheet (Microsoft Office, ver. 2018).
The wines of the 2019 vintage were evaluated after 3 mo of bottle aging. Due to the COVID-19 pandemic, the panel was not able to meet in person, so instead the wines were delivered to panelists’ houses, following the procedures for in home-use-test (HUT) sensory analysis, as outlined previously (Lawless and Heymann 2010). The panel that convened for the evaluation of the 2019 wines was composed of 10 individuals (n = 10; four females and six males), all of whom had previous experience in wine sensory analysis and were between 21- and 60-yr-old. Panelists were screened for visual disorders and bitterness sensitivity to PROP using the same procedure outlined for the 2017 and 2018 wines. The results obtained indicated that none of the panelists had pre-existing color deficiencies, and the panel was composed of 10% supertasters, 60% medium tasters, and 30% non-tasters.
Wine distribution to panelists was as follows. Each panelist received a tray of sixteen 50-mL Falcon tubes (Fisher Scientific) which were filled the same day of distribution and sparged with N2 upon filling. Twelve of the Falcon tubes were labeled with three-digit codes that were randomly generated and represented the three original replicates of the 10-, 12-, 14-mm, and unsorted wines, which were all randomly assigned. The raisin category was not included in the sensory analysis of the 2019 wines due to a lack of a sufficient amount of wine to effectively carry out the test. The other four Falcon tubes on the tray were composite samples of each repetition for a size category and were labeled A, B, C, and D, representing the 10-, 12-, 14-mm, and unsorted wines, all placed in randomized order. The day after the samples were gathered and dropped off to panelists, the panel had a virtual meeting to establish common sensory descriptors by tasting each composite sample. After deciding on 16 descriptors, ballots were created and dropped off to each panelist’s individual address. The descriptors selected by panel consensus included plum, raisin, red fruit, black fruit, spice character, pomegranate juice, cured meat, solvent, floral, sweetness, acidity, astringency, retronasal flavor, color saturation, red hue, and purple hue (Supplemental Table 1). Each page of the ballot packet listed one of the randomized codes which corresponded with one of the Falcon tubes and contained a line scale with the characteristics previously listed. Panelists were instructed to empty the content of each Falcon tube into ISO wine glasses and complete their blind tasting by themselves between 1000 and 1200 hr, before consuming lunch. Ballots were individually collected immediately after tasting, manually decoded in cm using a ruler, and input into an Excel spreadsheet (Microsoft Office, ver. 2018).
Statistical analyses
All treatments, including fermentations, were performed in triplicate. Differences among the mean values of each treatment for color and phenolic analysis were assessed through one-way analysis of variance (ANOVA), performed at 260 days postcrush. A series of two-way ANOVA were utilized to assess the individual effects of berry size category, vintage, and their respective interactions. In all cases, means were compared for significant differences using Fisher’s least significant differences test (p < 0.05). The coefficient of variation (CV) was calculated as the percentual ratio between the standard deviation and the mean for each parameter when appropriate. Unless otherwise indicated, all statistical analyses were conducted using XLSTAT (Add-insoft). GraphPad Prism software ver. 9.0 (GraphPad Software Inc.) was used for all graphical representations.
Sensory data were analyzed via principal component analysis (PCA) using the correlation matrix with no rotation, which was applied to each sensory data set in full, including the replicates, using R software ver. 3.4.0 (R Core Team 2021). Confidence ellipses indicating 95% confidence intervals were based on the multivariate distribution of Hotelling’s test for p < 0.05 and were constructed using the SensoMineR panellipse function of R, as described previously (Husson et al. 2005).
Results and Discussion
In this experiment, old-vine Zinfandel grapes from two different vineyard sources within the Paso Robles AVA of California were segregated over three consecutive vintages into four berry size categories, plus another category corresponding to unsorted (i.e., berries as they are collected and obtained after a standard destemming process, which served as control). These selected berry size categories were made into wine which was analyzed for chromatic, phenolic, and sensory parameters.
Berry basic chemistry and physical characteristics
The basic chemical composition of Zinfandel grapes as a function of berry size over three consecutive vintages is shown in Table 1, and the main physical and compositional characteristics of the berries of the different size categories are shown in Table 2. The CV for each of these parameters was also determined to understand the variability of a given parameter across the different berry size categories. Over the three vintages, TSS levels were consistently lower in the largest size category (i.e., 14-mm berries), followed by unsorted berries. In two of the three vintages, raisins had the highest TSS levels of all categories, reaching 36 Brix in the 2019 vintage and almost 30 Brix in the 2017 vintage. TSS levels had the highest CV for chemical composition, which was as high as 20% in 2019 (and 12% and 5% in 2017 and 2018, respectively) (Table 1). Conversely, pH values in the fruit were much less affected as a function of berry size. However, pH was affected as a function of vintage and vineyard site, with lower pH values in the 2018 vintage. Likewise, TA and malic acid levels were more affected by vintage, with higher values in the comparatively cooler 2019 vintage (Supplemental Table 2).
Vineyard source, harvest dates, and one-way analysis of variance (ANOVA) of fruit chemical composition in Zinfandel berries at harvest for each berry size category for the 2017, 2018, and 2019 vintages. Two-way ANOVA with p values for interaction between berry size category and vintage is also presented. Values represent the average of three independent samples of 30 berries each (n = 3). TSS, total soluble solids; TA, titratable acidity; YAN, yeast assimilable nitrogen; AVA, American Viticultural Area; CV, coefficient of variation.
Percentage distribution of fresh weight (FW) of fruit per berry class and one-way analysis of variance (ANOVA) of fruit physical composition in Zinfandel berries at harvest for each berry size category for the 2017, 2018, and 2019 vintages. Two-way ANOVA with p values for interaction between berry size category and vintage is also presented. Values represent the average of three independent samples of 30 berries each (n = 3). DW, dry weight; CV, coefficient of variation.
Overall, the percentage contribution of each size category to total FW decreased with decreasing berry size (Table 2). Thus, irrespective of the vintage and vineyard site, 14-mm berries accounted for 41 to 48% of the distribution on a FW basis. In contrast, raisins only accounted for 7 (in 2019) to 10% (in 2017) of the FW berry distribution (Table 2). Although the lowest contribution to total FW in the case of raisins is expected, this is directly tied to berry FW and pulp FW (Table 2). In addition to the inherent propensity of Zinfandel to produce raisins within a cluster, variations in the proportion of raisins are likely the result of vintage variation, especially in relation to extreme weather events. The 2017 vintage experienced four heat waves, with a total of 9 days with temperatures above 40°C (Supplemental Table 2), thereby resulting in a higher proportion of raisins in 2017. In contrast, the 2019 vintage did not experience heat waves, and 0 days of daily maximum temperatures above 40°C were registered.
Seed number per berry, skin and seed DW, and skin surface area were highest in 14-mm berries and lowest in 10-mm berries. Of these parameters, seed DW had the highest CV (up to 48% in 2018), but this was not the case for seed number per berry (CV between 25 and 27%). Previous research in Syrah (Walker et al. 2005) and Cabernet Sauvignon (Roby and Matthews 2004) has shown a direct relationship between berry size and seed number per berry, whereby larger berries typically had a higher number of seeds per berry. This is because seeds act as metabolic sinks of plant hormones, especially abscisic acid, which regulates berry growth (Wang et al. 2019). In accordance, our results suggest that in Zinfandel, larger berries have more seeds per berry, but also heavier seeds (Table 2). Because 14-mm berries had the lowest TSS levels but also more numerous and heavier seeds, this may imply that these larger berries are less ripe, physiologically and possibly aromatically, than smaller berries (Wang et al. 2019). The latter has implications for winemaking. Although the general extractability of seed tannins decreases over ripening, less ripe seeds may preferentially release tannins when the solvent is ethanol, as encountered during alcoholic fermentation (Bautista-Ortín et al. 2012). Thus, these less ripe seeds of larger berries may result in higher-than-expected tannin extraction into wine.
Large variations in berry FW were observed over the three vintages, with 2019 showing an almost 41% variation in berry FW, as reflected by the CV (Table 2). This contrasts sharply with reports of CV in the range of 25 to 30% in Merlot berries subjected to different irrigation schedules (Shellie 2010). Uniformly irrigated vines are expected to show even lower variations in berry FW than vines subjected to different irrigation schedules. For example, variations in berry diameter as low as 11% were observed for Cabernet franc (Pagay and Cheng 2010). This highlights the inherent heterogeneity of Zinfandel berry size, relative to other varieties.
Interestingly, the percentage of solids, i.e., skins and seeds (FW basis), showed relatively minor differences among berry size categories. In 2017, 10-mm berries had the highest percentage of solids (20%), but the same category had the lowest percentage of solids in 2018 and 2019 (12% and 11%, respectively). The desirability of smaller berry size for winemaking purposes is due to the putative higher proportion of solids in smaller berries, as solutes of sensory relevance are mostly located in the solid parts of the berry. In contrast, this research shows that smaller berries do not always have the highest percentage of solids, and thus the previous assumption appears unfounded. Indeed, seed and skin FW increased proportionally with berry size in all three vintages, as previously found elsewhere (Matthews and Nuzzo 2007). However, when skin FW and DW were compared and the percentage of water in skin tissue calculated, 10-mm berries had 63% water content over the three vintages, whereas 14-mm berries had 75% water content. Conversely, seeds showed no overall differences in percentage water (24% and 23% water for 10- and 14-mm berries, respectively).
Basic chemistry of the wines
The basic chemical composition of the wines made from each berry size category, along with the CV for each chemical parameter, is shown in Table 3. A direct relationship between initial TSS at harvest and final ethanol levels in the wines was observed, but some deviations occurred because alcoholic fermentation was not completed in the highest TSS categories (wines from 10-mm berries and raisins). Nonetheless, ethanol levels were highest in wines of the 10-mm berry category and lowest in wines of the 14-mm category. This resulted in an average difference in ethanol of 2.8% v/v between the wines of the 10- and 14-mm berry size categories. This result implies that berries of the smallest size categories within a cluster are the main drivers of the elevated ethanol levels in Zinfandel wines, which range from 16.5 to 17.8% (Riffle et al. 2022, Casassa et al. 2023).
Basic chemical composition and one-way analysis of variance (ANOVA) of Zinfandel wines made from different berry size categories for the 2017, 2018, and 2019 vintages. Two-way ANOVA with p values for interaction between berry size category and vintage is also presented. Values represent the average of three independent wine fermentation replicates per treatment (n = 3). TA, titratable acidity; CV, coefficient of variation.
Malolactic fermentation was completed in the unsorted and 12- and 14-mm berry size categories, but not in the wines of the 10-mm berries and raisins. This was likely due to the combined inhibitory effects on malolactic bacteria resulting from higher levels of malic acid (Table 1), and higher ethanol content (Table 3) in the wines of these small berry size categories.
Overall, berry physical parameters such as seed weight, seed number per berry, and pulp weight showed relatively large variations within berry size categories. However, variations of much lesser magnitude occurred as a function of wine chemical parameters (except for that due to unintended residual levels of sugar and malic acid in the wines), implying that the variation of certain berry physical traits does not always directly translate into differences of equal magnitude in the wines.
Chromatic characteristics of the wines
The wines were analyzed for selected parameters related to color composition as well as for parameters pertaining to sensory-relevant phenolic classes. The color composition of the wines in terms of total color (AU 420 + 520 + 620 nm) and hue (420/520 nm) (Glories 1984) is shown in Figure 1. Wines were assessed at day 14 (at pressing) and after 3 mo of bottle aging (day 260). Uniquely, the chromatic and phenolic composition of the 2017 wines was reassessed after 6 yr of bottle aging.
Evolution of wine color and wine hue at pressing (day 14) and after 3 mo of bottle aging (BA) in Zinfandel wines produced from five different berry size categories. The wines of the 2017 vintage were reassessed after 6 yr of BA. Different letters at 3 mo BA indicate significant differences for Fisher’s least significant differences test (p < 0.05).
In most wines, overall chromatic differences observed at pressing were maintained after 3 mo of bottle aging. These differences followed a recognizable trend in that color intensity in the wines was inversely related to berry size. Wines from raisins and those made from 10-mm berries had the highest wine color. Conversely, wines made from the largest berry size category (14 mm) had the lowest color. This result agrees with that from a previous study in which Cabernet Sauvignon berries were sorted into three berry size categories. Out of several parameters measured in the resulting wines, wine color (measured with the same methodology of the present study) was the wine parameter most affected by berry size, whereby the smallest berries produced the deepest colored wines (Gil Cortiella et al. 2020). In larger berries, dilution occurs in a manner proportional to the decrease of the surface-to-volume ratio (Matthews and Anderson 1989, Hardie et al. 1996). Although the skin surface area was lowest in 10-mm berries (and accordingly, highest in 14-mm berries), the skin surface-to-berry FW ratio was highest in 10-mm berries (3.12) and lowest in 14-mm berries (2.32). Herein we show that wine color, which is largely regulated by skin-derived anthocyanins (Casassa and Harbertson 2014), is a function of comparatively lower berry volume and lower skin water percentage (thus leading to concentration), rather than of only higher skin surface area (Figure 2). The implication of this finding is that with increasing berry size, the pulp (mesocarp) increases proportionally more than the surface of the pericarp (skin), resulting in a comparatively lower surface-to-berry weight ratio, and thus potentially causing dilution effects (Figure 2). Conversely, smaller berries have more skin surface in proportion to their volume.
Relationship between percentage (%) of water on skins and seeds, the skin surface-to-berry fresh weight (FW) ratio in grapes, and color of wines of the 10-, 12-, and 14-mm berry size categories. Wine color, tannins, and polymeric pigments correspond to the color of the wines after 3 mo of bottle aging. Values correspond to average values (followed by the standard error of the mean) of three vintages.
Hue (AU 420/520 nm) was also determined in the finished wines after 3 mo of bottle aging, and further after 6 yr of bottle aging in the 2017 wines. Wines from the 10-mm berries had the lowest hue (less brown, aging-related tones), and wines from the 14-mm berries had the highest hue (more aging-related hues). This shows that wines from the 10-mm berries display higher color saturation and a deeper, darker, more purple color than wines from larger berry size categories.
Phenolic composition of the wines
Selected phenolic classes across the three consecutive vintages under study in the wines of the different berry size categories throughout winemaking, including anthocyanins (responsible for wine color) and tannins (responsible for mouthfeel properties such as astringency), are shown in Figure 3. In addition, the detailed polymeric pigment composition of the wines segregated into SPPs (which include low molecular weight pigments such as pyranoanthocyanins) and LPPs (higher molecular weight pigments able to elicit astringency) is shown (Figure 4). Anthocyanins registered peak concentration at pressing (although this does not preclude the possibility that higher anthocyanin concentrations may have been reached earlier during fermentation), then experienced a decrease after 3 mo of bottle aging. At this time and even after 6 yr of bottle aging, anthocyanins were consistently higher in the wines from the 2017 vintage made from 12- and 14-mm berry size categories. In contrast, wines made from 10-mm berries and raisins had comparatively lower levels of free anthocyanins. The following incongruence emerges as apparent: although anthocyanins are primarily responsible for wine color, wines from the 12- and 14-mm berries, which were higher in anthocyanins, did not show higher color saturation (Figure 1). Lower color saturation in the wines of the larger berry size categories can be due to comparatively high wine pH, as in the case of the wines of the 12- and 14-mm berry size categories in 2018 and 2019. The discrepancy between higher anthocyanin content and lower color in wines from 14-mm berries can also be explained by the fact that wine color is a complex phenomenon that depends not only on anthocyanins, but also on tannins, polymeric pigments, and other metabolites, including noncovalent reactions such as copigmentation (Boulton 2001). Supporting the previous statement, a study that looked at the effects of berry size on wine composition linked higher absorbances at 420, 520, and 620 nm with the total polyphenolic content of the wines (measured as total polyphenol index) (Gil Cortiella et al. 2020), a correlation also observed herein and further addressed below. Thus, lower color saturation in wines from 14-mm berries may have also been linked to an overall low total phenolic content, as well as to a low polymeric pigment content.
Evolution of anthocyanins, tannins, total phenolics, and non-tannin phenolics at pressing (day 14) and after 3 mo of bottle aging (BA) in Zinfandel wines produced from five different berry size categories. The wines of the 2017 vintage were reassessed at day 2200 (6 yr of BA). Different letters at 3 mo BA indicate significant differences for Fisher’s least significant differences test (p < 0.05).
Evolution of small and large polymeric pigments (SPP and LPP, respectively) at pressing (day 14) and after 3 mo of bottle aging (BA) in Zinfandel wines produced from five different berry size categories. The wines of the 2017 vintage were reassessed after 6 yr of BA. Different letters at 3 mo BA indicate significant differences for Fisher’s least significant differences test (p < 0.05).
If anthocyanin content in wine was a sole function of skin surface (as can be derived by the higher levels of free anthocyanins in wines from 14-mm berries compared with those from 10-mm berries), then the anthocyanin content in 10-mm wines at pressing could be predicted by calculating proportionally the corresponding amount of wine anthocyanins. When these calculations were performed, we found that the predicted wine anthocyanin content based on the decrease in skin surface is less in 10-mm berries than what the wines actually show at pressing. It therefore follows that on average over the three vintages, the skins of 10-mm berries were 52% more concentrated in anthocyanins than the skins of 14-mm berries, based on anthocyanin levels recorded at press.
Tannin concentration in the finished wines was statistically higher in the wines of the smaller berry size category (i.e., raisins and 10-mm berries). Particularly noteworthy was the generally high tannin content of the wines made from the raisin berry size category. The exception was 2019, in which raisins reached the highest sugar levels (36 Brix; Table 1) and accounted for the lowest percent distribution of FW (Table 2) but yielded wines with lower tannin concentration. Conversely, the lowest tannin concentration (close to the limit of detection of the protein precipitable tannin method, i.e., 50 mg/L) was observed in the wines of the 14-mm berry size category. Tannins are located in both seeds and skins, with generally higher tannin concentration in seeds than in skins (Harbertson et al. 2002). Although 14-mm berries had the highest average number of seeds per berry and the heaviest seeds (Table 2), they also showed the lowest surface area-to-berry FW ratio, as previously discussed. This lower skin surface-to-berry FW ratio in 14-mm berries may also explain the lower tannin content in the wines made from this berry size category (Figure 2), and possibly, the lower concentration of skin tannins in these berries (as with anthocyanins). Indeed, previous research on a Croatian variety, Plavac mali, genetically related to Zinfandel, reported tannin concentration approximately three times higher in skins relative to seeds (Ćurko et al. 2014). From a winemaking perspective, extended maceration for 30 days, known to preferentially favor the extraction of seed-derived tannins in other varieties, when applied to Zinfandel seems to have no impact on enhancing the effects of extracted tannins (Casassa et al. 2019), suggesting that Zinfandel berries may be high in skin tannins and low in seed tannins. These observations imply that for Zinfandel, tannin extraction may be more of a function of skin mass and surface rather than seed number and seed weight. It thus follows that for Zinfandel wines, skin-derived extraction of tannins may be more of a relevant factor than seed-derived extraction.
Total phenolics (i.e., iron-reactive phenolics) include tannins, flavan-3-ols, and flavonols, but exclude monohydroxylated phenols and anthocyanins (Figure 3). In all instances, total phenolics were higher in wines from the raisin berry size category. In 2018 for example, the concentration of total phenolics in the wines made from raisins was nearly three times higher than in the wines made from unsorted berries (2317 mg/L and 823 mg/L, respectively) (Figure 3). The other berry size categories showed less variation in total phenolics, although the wines from the 10-mm berries tended to have higher concentrations. A very similar trend was observed in the case of non-tannin phenolics. Non-tannin phenolics represent the amount of total phenolics minus the amount of tannins, effectively estimating the concentration of monomeric flavan-3-ols and low molecular weight tannin dimers (Figure 3). The wines from raisins had the highest concentration of non-tannin phenolics and the wines from 14-mm berries and unsorted berries had the lowest. Taken collectively, these results suggest that raisins, primarily, and 10-mm berries in Zinfandel are particularly low in anthocyanins but high in tannin and other non-anthocyanin phenolics, which may explain the higher color saturation in their resulting wines.
The polymeric pigment composition in the wines is shown in Figure 4. Polymeric pigments were segregated into SPPs and LPPs. In all cases, polymeric pigments increased from pressing to 3 mo of bottle aging. Polymeric pigments in the 2017 wines remained stable in the 10-mm berry size wines and decreased in the wines made from raisins after 6 yr of bottle aging.
In all three vintages, wines of the 10-mm berry size category and those of raisins were consistently higher in polymeric pigments. The lowest level of polymeric pigments was observed in the wines of the 14-mm berry size category. As with other phenolic classes, the polymeric pigment content of the resulting wines was inversely related to berry size, whereby higher contents were observed in the wines from the smallest berry size categories (Figure 2). This higher polymeric pigment content in the wines of the smaller berry size categories may also contribute to the higher color saturation in their respective wines. It may also explain the observed lower content of free anthocyanins in the wines of the smaller berry size categories (Figure 3).
Sensory composition of the wines
To further connect and relate the effects of berry size to selected sensory parameters, the wines of the three vintages were analyzed via descriptive analysis (Figure 5). Sensory results were plotted using PCA coupled with confidence ellipses. The ellipses represent empirical descriptions of the variability of the sensory evaluations, implying that a lack of overlap between confidence ellipses corresponding to two or more given wines indicates statistical support to confirm that these wines are significantly different in their sensory features (Husson et al. 2005). While some vintage-specific effects on the sensory composition of the resulting wines were observed, several commonalities emerged that allowed for the following generalizations.
Principal component analysis of descriptive sensory data generated by trained panels. Data from Zinfandel wines were produced from five different berry size categories. A) 2017 vintage, B) 2018 vintage, C) 2019 vintage. Confidence ellipses indicate 95% confidence intervals.
The most distinctive wines were those made from 10-mm berries and raisins (especially in the 2017 and 2018 vintages), whereby their respective ellipses did not overlap and clearly shifted away from the other berry size categories. Wines made from the 10-mm berries had the highest astringency (in 2017 and 2019), a higher and more prominent spicy character (in all three vintages), and the highest purple hue and color saturation (in 2017 and 2019). Wines made from raisins had the highest sweetness in 2017 and 2018 (in accordance with measurable residual sugar levels in these wines) (Table 3) and were defined by aromas of cough syrup and black fruit. In 2018, wines made from raisins had the highest perceived astringency, visual color saturation, and purple hue. The higher color saturation perceived in the wines of the smaller berry size categories can now be unequivocally related to the enhanced overall phenolic composition (Figure 3) and polymeric pigment content (Figure 4) in these wines.
The wines from the 12- and 14-mm berry size categories and the wines made from unsorted berries were all placed in the negative dimension of each PCA plot for all three vintages. The sensory features of the wines of unsorted berries overlapped with those of 12-mm berries in 2017, 12- and 14-mm berry wines in 2018, and 14-mm berry wines in 2019. The overlapping sensory features of the wines from unsorted and 14-mm berries is consistent with 14-mm berries representing close to 50% of the distribution of berry FW (Table 2). Notwithstanding, this relationship is not expected to be linear, as other berry size categories that are proportionally less abundant may provide flavors and aromas that blend with, or override, those of 14-mm berries.
Across the three vintages, wines from unsorted-, 12-, and 14-mm berries had higher perceived acidity, red fruit aromas (in 2017 and 2019), and earthy aromas, and overall had much less aromatic intensity than wines made from raisins and 10-mm berries (Figure 5). From a chemical viewpoint, wines from 12- and 14-mm berries had comparatively lower TA (Table 3) than wines from 10-mm berries, suggesting their perceived acidity should be lower than that of wines from 10-mm berries. However, the perception of sourness, or acidity, increases as alcohol content decreases (de-la-Fuente-Blanco et al. 2024), as is well-known in dealcoholized wines (Kumar et al. 2024), which typically display enhanced sourness, suggesting a masking effect of ethanol content on perceived acidity. Therefore, higher perceived acidity on the 12- and 14-mm wines may be due to their comparatively much lower ethanol content (as much as 4.8% less in wines of 14-mm berries relative to wines of 10-mm berries in 2017), causing the existing acids in the wines to be more easily perceived. Nevertheless, these results suggest that these berry size categories may provide acidity as well as red fruit character to the final sensory composition of the wines.
Overview of berry size effects on selected grape and wine parameters
The relative percentage increase or decrease of selected parameters in the wines of each berry size category is shown in Figure 6, whereby the baseline for percentage calculation were values recorded in the unsorted berries and/or their respective wines.
Percentage increase or decrease of selected chemical and sensory attributes of Zinfandel wines produced from four different berry size categories. The shown selected attributes that were recorded in unsorted berries and wines were used as a baseline to calculate percentages in the wines of each berry size category. TSS, total soluble solids.
In relative terms, TSS variations in the fruit and the final alcohol content of the wines were of lesser magnitude than variations in phenolics and color and some selected sensory features. Wine color showed increases ranging from 75% to upward of 100% in wines of the 10-mm berries and raisins. Conversely, up to 50% less color was observed in the wines made from the 14-mm berries in 2017. Anthocyanins, however, were ~30% higher (in 2017) and upward to 75% higher (in 2019) in the 12-mm berry wines. Tannins in the wines from raisins increased by 450% in 2018 (and up to 175% in 2017) relative to wines from unsorted berries. Interestingly, comparatively lower tannins levels were found in wines from raisins in 2019, but a causative reason for this is only speculative because these wines showed the lowest ethanol levels and did not complete alcoholic fermentation. Percent increases in polymeric pigments showed similar trends to those observed for tannins. For example, in 2019, polymeric pigments were up to 130% higher in wines made from the 10-mm berry size category. Visual color generally mirrored measured color (Figure 1), with consistent increases ranging from 50 to 55% (in 2018 and 2019), and up to 75% in 2017, in the wines made from 10-mm berries.
Consistent with previous results gleaned from the sensory composition of the wines (Figure 5), wines made from the 10-mm berries showed the greatest increases in spicy aroma. Similarly, the perception of dried fruit was ~50% higher in the wines from raisins in 2017 and 2018. Wines of the 14-mm berry size category, on the other hand, had up to 50% less dried fruit character than that registered in the wines from unsorted berries. Sweetness perception was a function of residual levels of sugars in the finished wines. Finally, astringency was consistently higher in the wines made from 10-mm berries and raisins (up to 50% higher in 2017). Those percentual increases, albeit significant, are of lesser magnitude than those previously observed variations in tannin content in the wines from 10-mm berries and raisins. Indeed, the tannin content of the latter wines was the factor that, on a percentage basis relative to unsorted berries, varied the most among the wines of the different berry size categories (Figure 6).
Conclusion
As a varietal of historical relevance in California, there has been continued interest in understanding the reasons that underpin the drastic variations in the size and ripening of Zinfandel berries, as well as that of the high heterogeneity of its physical and chemical traits in the vineyard. The implication is that such heterogeneity will ultimately affect the chemical and sensory composition of its wines. Over three consecutive vintages in the Paso Robles AVA of the Central Coast of California, this study documents for the first time the physical and chemical traits of differently sized (14-, 12-, and 10-mm berries; raisins; and unsorted berries) Zinfandel berries, and the effect of these sizes on the chemical and sensory composition of their resulting wines.
While the established belief was that a higher percentage of berry solids (i.e., skins and seeds) could be used as a proxy to predict wine quality, we showed that smaller berries rarely have the highest percentage of solids. Nonetheless, these small berries do result in unique chemical and sensory features, some of which are associated with higher quality wines. Smaller Zinfandel berries show proportionally higher seed FW and skin FW relative to their volume. Because berries are spherical, which maximizes the surface-to-volume ratio, it follows that smaller berries should maximize such a ratio. However, berry skins have thickness; thus, more precisely, the smaller 10-mm Zinfandel berries maximize the surface-to-berry FW ratio. In other words, smaller berries have more skin surface (and skin content) in proportion to their FW, and conversely, larger berries have less skin surface because their FW increases proportionally more than the increase in skin surface. Considering all three vintages, in going from 14- (largest berry size category) to 10-mm berries (smaller berry size category), we calculate that berry FW decreases by an average of 61%, whereas skin surface decreases by 43%. That is, the FW decreases proportionally less with decreasing berry size than the solids. In addition, the skins of 10-mm berries are more concentrated in solutes (e.g., anthocyanins and tannins) than the skins of larger berries, as they hold less water than that of the skins from larger berries. The surface-to-FW ratio and the concentration of this surface are the two key parameters related to the final concentration of solutes of sensory relevance in wine. As a result of this ratio, wines made from 10-mm berries had enhanced chromatic characteristics, higher perceived color, higher content of polymeric pigments (with desirable implications for mouthfeel), higher tannin concentration and astringency, and overall higher aromatic intensity than wines made from larger berries. Furthermore, it is likely that tannin extraction and retention in Zinfandel wine is regulated by skin- and not seed-derived tannins, and the overall tannin concentration may be low in Zinfandel seeds. This was evidenced by the fact that larger berries had the highest number of seeds per berry and the heaviest seeds, yet the lowest tannin content, in the wines.
The documented proportion of raisins in the present study ranged from 7 to 10% of the FW distribution and reached up to 36 Brix, which poses a practical problem for winemakers. This is because raisins typically do not release their content until after alcoholic fermentation is well underway. The result of this is an underestimation of TSS levels measured in the fermentor upon crushing.
It was also found that raisins and 10-mm berries had higher ethanol levels in the finished wines. This finding has profound implications for winemaking, whereby optical sorters could be employed to sort out the berries of the smaller size categories, particularly raisins, to produce wines with higher ethanol content. Winemakers targeting lower alcohol levels and less ripe, raisined flavors in the resulting wines should therefore aim to selectively discard raisined Zinfandel berries. However, more classic styles of Zinfandel could be targeted by retaining smaller berries and even raisins. The present work showed that 10-mm berries are responsible for the trademark spicy character of Zinfandel wines, whereas raisins may provide aromas and flavors related to dried fruit and/or cooked fruit, including aromas of raisins, cough syrup, and black fruit, and possibly enhanced perceived sweetness. These sensory features are considered to be part of the distinctive sensory profile of California Zinfandel wines; high ethanol levels and smaller berry size categories contribute to these features.
Larger berry size categories generally produced wines with higher perceived acidity and red fruit character, and lower spicy and dried-fruit characteristics, which were related to lower TSS levels at harvest in larger berries. Thus, wines made from larger berries (i.e., 12- and 14-mm berries in the present study) will display comparatively lower ethanol content and unique sensory characteristics, likely to be perceived as more fruit-forward and less “ripe”. These features of Zinfandel wines made from larger berries may be stylistically desirable. When determining the final sensory composition of Zinfandel wines, TSS level may be a good predictor of the sensory features of larger berries on an individual berry-by-berry basis within a cluster. Therefore, berry sorting on Zinfandel may offer the possibility of creating wines with distinct sensory features.
Supplemental Data
The following supplemental materials are available for this article in the Supplemental tab above:
Supplemental Table 1 Color, aroma, and mouthfeel attribute standard composition and specifications for descriptive analysis of Zinfandel wines.
Supplemental Table 2 Growing degree days (GDD), Winkler regions, maximum temperatures and heat waves, and annual and seasonal precipitation in the 2017, 2018, and 2019 growing seasons for Atascadero, CA.
Supplemental Figure 1 Photograph of the berry sorting device, with different trays and sieve diameters in mm.
Data Availability Statement
The data underlying this study are available on request from the corresponding author.
Footnotes
The Dusi family is generously acknowledged and thanked for donation of fruit and assisting with the logistics of harvest. Karl Wicka and Turley Wine Cellars are also thanked for donation of fruit and harvest logistics.
Casassa LF, Palmer NR, Donegan AV, Forrey SK, Postiglione DA, Catania AA and Dodson Peterson JC. 2025. Influence of berry size on California-grown Zinfandel grapes and wines. Am J Enol Vitic 76:0760017. DOI: 10.5344/ajev.2025.25004
By downloading and/or receiving this article, you agree to the Disclaimer of Warranties and Liability. If you do not agree to the Disclaimers, do not download and/or accept this article.
- Received January 2025.
- Accepted April 2025.
- Published online July 2025
This is an open access article distributed under the CC BY 4.0 license.
References
Vol 76 Issue 2
Jump to section
Related Articles
Cited By...
More from this TOC section
Similar Articles
